Hydraulic Loss–Based Genetic Optimization and CFD Assessment of a Reversible Centrifugal-Pump Impeller Operating in Pump and Turbine Modes

This study presents a computationally inexpensive preliminary sizing and parametric-design procedure for a centrifugal-pump impeller intended for exploratory operation in pump and turbine modes. A hydraulic loss model was coupled with a real-valued genetic algorithm to screen a restricted design space without embedding CFD simulations in the evolutionary loop. The selected geometry was reconstructed and comparatively assessed with steady RANS simulations in OpenFOAM 12 using the k-omega SST turbulence model. Relative to the reference parametric geometry, CFD results at the prescribed conditions showed efficiency increases of 7.48 percentage points in pump mode and 11.05 percentage points in turbine mode. The results do not validate a dual-mode optimization method or establish performance over an operating range. They show that, within the selected correlations, parameter choices, and passive geometric constraints, a combined geometric modification can improve the numerical response at the investigated conditions. CFD and experimental assessment over complete performance curves are required before broader design claims are made.